I create some models like this using a nested tidyr dataframe:
set.seed(1) fits <- tribble(~group, ~colA, ~colB, ~colC, sample(c("group1"